Many forms of throttling valve are found within the field. Viraraghavan describes a valve in patent '530(cited in the references) that that is typical of valves that may, at first, may be deemed to be satisfactory for these applications. However, as is typical in many valve designs the stem must be sealed with o-rings. This introduces both small cracks and crevasses, as well as another material(the o-ring) that may not be compatible with the process fluid. These problems can be solved with diaphragm valves. Linder, in patent '086(sited in the references) describes one such approach. Here a metallic actuation shaft is buried in the flexible diaphragm such that it is susceptible to corrosion from ionic migration from some process fluids. Secondly, this valve contains a large number of parts of metallic content as well as a comparative large volume of fluid should the primary diaphragm fail. Generally valve designs described in the art attack one or two of the problems described, but none approach the whole problem of safely handling highly toxic or corrosive fluids while providing a smooth continues laminar flow in a free draining, low captive volume device.
The throttling mechanisms described in the art contain sharp corners and short flow paths to control the flow. Many fluids are both hazardous and suffer detrimental effects to sheer effects when flowing around such obstructions. The slurries used in the semiconductor industry are known for their sensitivity to shear. It is generally known in fluid mechanics, that laminar flow is achieved for Reynolds numbers of less then 2000. Where the Reynolds number is computed as follows:Re=Dvρ/μe 
Where D is the height of the flow channel, v the fluid velocity, ρthe density, and μe the fluid viscosity. Furthermore, the pressure drop, ΔP, along the channel is expressed differently for laminar flow that than for turbulent flow which occurs and at higher Reynolds numbers.For Laminar flow: ΔP=K(μLv/D2)For turbulent flow: ΔP=K′(UD)(v2)
Where K and K′ are constants of proportionality that depend on the engineering units selected. It can be clearly seen that the pressure drop, hence, the throttling effect on the flow stream is a linear relationship to the fluid velocity for laminar flow and a velocity squared relationship for the turbulent flow case. Obviously the ideal case is a linear relationship of pressure loss to flow velocity across the throttling range. A less desirable situation is the square law response where the response of the valve continuously varies over the throttling range. Any configuration that results in the flow response being linear over a portion of the range and transitioning to square law at some point will have a detrimental effect on the ability of the controller to smoothly regulate the flow. In additions addition, since both relationships contain the length factor L in the numerator it is desired to have some length to the throttling path.
In order to provide a throttling effect it is necessary to control the opening of the fluid channel. Many means are described in the art. These are typically motors, solenoids, compressed air, or manual adjustment devices. The applications envisioned herein all contain a flow meter to measure the flow and advanced electronics to position the valve to achieved achieve the desired mass or volumetric flow rate. Thus, the position of the valve is not so much of interest as is the results measured by a suitable flow meter. However, in such feed back arrangement the valve undergoes continuous repositioning, and therefore must have a reliable drive mechanism with little or no backlash. Furthermore, as the fluids may be under pressures up to 100 psi the drive mechanism must operate with a wide range of axial loads. Generally drive mechanisms described in the references were judged as incapable of continuous repositioning, or would exhibit excessive backlash detrimental to smooth fluid control. A second feature found to be overlooked in the current art is the ability to remove the drive mechanism and inspect it without the risk of opening the fluid path.
Reviewing the references cited we find that none of them attempt to solve the full range of the problems posed herein.